Applications of Neodymium Isotopes to Ore Deposits and Metallogenic Terranes; Using Regional Isotopic Maps and the Mineral Systems Concept

Abstract

Although radiogenic isotopes historically have been used in ore genesis studies for age dating and as tracers, here we document the use of regional- and continental-scale Sm–Nd isotope data and derived isotopic maps to assist with metallogenic interpretation, including the identification of metallogenic terranes. For the Sm–Nd system, calculated Nd model ages, which are time independent, are of most value for small-scale isotopic maps. Typically, one- or two-stage depleted mantle model ages (T_DM, T_2DM) are used to infer age when the isotope characteristics of the rock were in isotopic equilibrium with a modelled (mantle) reservoir. An additional advantage is that Nd model ages provide, with a number of assumptions, an estimate of the approximate age of continental crust in a region. Regional- and continental-scale Nd model age maps, constructed from rocks such as granites, which effectively sample the middle to lower crust, therefore, provide a proxy to constrain the nature of the crust within a region. They are of increasing use in metallogenic analysis, especially when combined with a mineral systems approach, which recognizes that mineral deposits are the result of geological processes, at a scales from the ore shoot to the craton. These maps can be used empirically and/or predictively to identify and target large parts of mineral systems that may be indicative, or form part of, metallogenic terranes. Examples presented here include observed spatial relationships between mineral provinces and isotopic domains; the identification of old and/or thick cratonic blocks; determination of tectonic regimes favorable for mineralization; identification of isotopically juvenile zones that may indicate rifts or primitive arcs; recognition of crustal breaks that define metallogenic terrane boundaries or delineate fluid pathways; and, as baseline maps. Of course, any analysis of Sm–Nd and similar isotopic maps are predicated on integration with geological, geochemical and geophysical information data. In the future, research in this area should focus on the spatial and temporal evolution of the whole lithosphere at the province- to global-scales to more effectively targeting mineral exploration. This must involve integration of radiogenic isotopic data with other data, in particular, geophysical data, which has the advantage of being able to directly image the crust and lithosphere and being of a more continuous nature as compared to invariably incomplete isotopic data sets.

1 Introduction

Samarium-Neodymium (Sm–Nd) isotopes have commonly been used in economic geology studies, not just for geochronology, but also as tracers that provide constraints on mineralizing and geological processes. As a tracer Sm–Nd data constrain the age and source of mineralization, fluid pathways, metal sources and mineralizing processes. There is extensive literature regarding such use of radiogenic isotopes; review papers include Tosdal et al. (1999), Lambert et al. (1999), and Ruiz and Mathur 1999). Recently, the use of the isotopes, such as Sm–Nd, has extended to regional-scale metallogenic analysis, importantly utilizing isotopic data from both mineralized and unmineralized rocks (e.g., Champion 2013; Champion and Huston 2016). This builds on recognition that regional-scale isotopic data can assist with identification of metallogenic terranes (Zartman 1974; Farmer and DePaolo 1984; Wooden et al. 1998). This approach has benefitted from the large quantity of isotopic data now available and graphical software, which have enabled the rapid and repeatable generation of isotopic maps at regional to continental scales. Examples for the Sm–Nd system include studies by Cassidy et al. (2002), Champion and Cassidy (2007, 2008), Champion (2013), Huston et al. (2014), Mole et al. (2013), and Wu et al. (2021). Champion’s (2013) Australia map was the first Sm–Nd isotopic map produced for an entire continent, undertaken at the Australian continent-scale.

Following pioneering studies such as Zartman (1974), the usefulness of isotopic distribution patterns for metallogenesis has become increasingly recognized. Although initially driven by empiricism (Zartman 1974; Wooden et al. 1998; Cassidy and Champion 2004; Huston et al. 2005, 2014), these studies are now strongly influenced by the mineral systems concept (Wyborn et al. 1994), which recognizes that mineral deposits, although small, result from geological processes that occur, and can be mapped, at larger scales (Fig. 1: e.g., McCuaig et al. 2010; Hronsky et al. 2012; Huston et al. 2016). It is the larger camp- to continent-scale scales that regional Sm–Nd isotopic maps have most use, for example to infer continent- to province-scale lithospheric structures and architecture. This places both known deposits in context, but also allows targeting of undiscovered deposits by constraining geodynamic settings, fluid and magma pathways, and energy and metal sources.

Fig. 1

Mineral system model and examples. a Mineral system model illustrating the range of processes required to produce a mineral deposit. Such processes operate at a range of spatial and temporal scales (modified from Huston et al. 2016). b Cartoon illustrating potential processes that may be important for a specific mineral deposit. Note the large range of scales relative to the scale of the deposit (modified extensively from Australian Academy of Science 2012)

This contribution, therefore, details the Sm–Nd isotopic system, which has been use for over 40 years (Farmer and DePaolo 1983, 1984; Bennett and DePaolo 1987), and for which large regional isotopic datasets are available (e.g., Fraser et al. 2020). It discusses the uses and implications Sm–Nd isotopes and isotopic maps to different mineral systems, focusing on the craton- to district-scale. Examples include identifying lithospheric boundaries that commonly control sites of mineralization, determining fertile metallogenic provinces using isotopic characteristics, and establishing rock types essential to mineralization by their isotopic signature. We present general principles of Sm–Nd system; identify time-independent isotopic variables; and show how these variables can be used to generate isotopic maps useful to metallogenic studies. Much of the discussion and usage presented here applies to other isotopic systems, especially Lu–Hf (Osei et al. 2021; Waltenberg 2023) and U–Th–Pb (Champion and Huston 2016; Huston and Champion 2023).

2 The Samarium-Neodymium Isotopic System

As there are many reviews written on isotope systematics (e.g., Faure 1977; Dickin 1995) in general and on the Sm–Nd system in particular (e.g., DePaolo 1988; Champion and Huston 2016), only a brief introduction is provided here.

The equation describing the evolution of the Sm–Nd system through time can be written as follows:

$$\left( {^{143} {\text{Nd}}/^{144} {\text{Nd}}} \right)_{\left( {{\text{now}}} \right)} = \left( {^{143} {\text{Nd}}/^{144} {\text{Nd}}} \right)_{\left( {{\text{initial}}} \right)} + \left( {^{147} {\text{Sm}}/^{144} {\text{Nd}}} \right)_{\left( {{\text{now}}} \right)} *\left( {{\text{e}}^{\lambda {\text{t}}} - 1} \right)$$

(1)

where: λ = the decay constant for ¹⁴⁷Sm to ¹⁴³Nd (6.54 × 10^–12 yr^-1: Dickin 1995), (now) = abundance of the isotope as measured in the present day, (initial) = abundance of the isotope at time t in the past, and t = the age of the last isotopic disturbance rock in years. The isotopic disturbance age can record the magmatic, metamorphic, or mineralization age of the rock.

For tracer applications, often used in metallogenic studies, the initial parent/daughter ratio (initial ratio), which is commonly more useful, can be determined as follows:

$$\left( {^{143} {\text{Nd}}/^{144} {\text{Nd}}} \right)_{\left( {{\text{initial}}} \right)} = \left( {^{143} {\text{Nd}}/^{144} {\text{Nd}}} \right)_{\left( {{\text{now}}} \right)} - \left( {^{147} {\text{Sm}}/^{144} {\text{Nd}}} \right)_{\left( {{\text{now}}} \right)} *\left( {{\text{e}}^{\lambda {\text{t}}} - 1} \right)$$

(2)

Initial ¹⁴³Nd/¹⁴⁴Nd ratios can be reported as deviations (ε_Nd, reported in parts per 10,000) from a chondritic earth reference model (CHUR = Chondritic Uniform Reservoir: DePaolo and Wasserburg 1976):

$$\varepsilon_{{\text{Nd}}} = 10000*\left[ {\left( {^{143} {\text{Nd}}/^{144} {\text{Nd}}} \right)_{{\text{Sample}}\left( {\text{T}} \right)} -\left( {^{143} {\text{Nd}}/^{144} {\text{Nd}}} \right)_{{\text{CHUR}}\left( {\text{T}} \right)} } \right]/\left( {^{143} {\text{Nd}}/^{144} {\text{Nd}}} \right)_{{\text{CHUR}}\left( {\text{T}} \right)}$$

(3)

Although both initial ¹⁴³Nd/¹⁴⁴Nd ratios and ε_Nd values are used, the latter has the advantage on ε_Nd—time plots where depleted mantle and crustal reservoirs have distinct signatures.

3 The Samarium-Neodymium Geochemical System

The behavior of isotopic systems and the interpretation of isotopic data, is mostly a function of the relative geochemical behavior of the parent and the daughter isotopes. For the Sm–Nd isotopic system, both parent and daughter isotopes are lanthanides, or rare earth elements (REE), that all have very similar geochemical properties. Consequently, Sm and Nd have similar, predictable, geochemical behavior in most geological processes, resulting in very minor fractionation and minimal variation in Sm/Nd for common crustal rocks (Table 1). The fractionation that does occur is mostly related to lanthanide contraction, whereby lanthanides with higher atomic number have progressively smaller atomic radii. Hence, Nd is more incompatible than Sm, and Nd is preferentially concentrated in the melt. As a result, processes such as partial melting and fractional crystallization result in higher levels of the light REE and lower Sm/Nd ratios in the more differentiated end-members (DePaolo 1988). As a consequence, the Earth’s crust is enriched in the light REE and also has lower Sm/Nd than complementary depleted mantle reservoirs. Differences in Sm/Nd of mantle and crustal reservoirs result in divergent isotopic behaviors in these two reservoirs with time (DePaolo and Wasserburg 1976; Bennett and DePaolo 1987; DePaolo 1988; Table 1; Fig. 2).

Table 1 Average Nd and Sm concentrations (in parts-per-million) and Sm/Nd ratios of mantle and crustal reservoirs

Fig. 2

Epsilon Nd (ε_Nd) through time and model age plots (modified after Champion 2013). a Time-integrated behaviour of ε_Nd in continental crustal reservoirs versus the complementary depleted mantle reservoir. Schematic REE plots illustrate the idealised change in Sm/Nd ratio (normalised to chondrite) between the reservoirs. CHUR equals Chondritic Uniform Reservoir. b Nd model ages. Single stage Nd model ages (T_CHUR, T_DM) assume no fractionation of Sm/Nd (= measured ratio). The intersection of the sample evolution curve with the mantle evolution curve (chondritic (CHUR) or depleted mantle (DM)), defines the model age. Two-stage model ages (T_2DM) assume a change in Sm/Nd at some point in the crustal history of the protolith. For felsic magmatic rocks this is typically the magmatic age. For ages older than the magmatic age a model ¹⁴⁷Sm/¹⁴⁴Nd ratio is used, typically that of average continental crust. To calculate ε_Nd, values of ¹⁴⁷Sm/¹⁴⁴Nd and ¹⁴³Nd/¹⁴⁴Nd have to be assigned for the DM and CHUR reservoirs. DM values used here are 0.2136 and 0.513163, respectively

The long half-life of ¹⁴⁷Sm (10⁶ Gyr) and the geochemistry of the lanthanides make the Sm–Nd system ideal for documenting crustal development. Because of the long half-life, significant changes in the Nd isotopic signature require very long time frames, making the Sm–Nd system very useful to identify ancient crust. The isotopic effect of addition of isotopically juvenile material, through the incorporation of a juvenile mantle component in granitic magmas, depends mostly on the volume of juvenile material added. As REEs are enriched in granitic rocks relative to mantle material, the Nd isotopic signature of granites is often insensitive to addition of significant amounts of such mantle material. For example, Kirkland et al. (2013) show that the isotopic signatures of Proterozoic REE-enriched granites in the Musgrave Province of central Australia have been insensitive to the addition of as much as 85% mantle input. While an extreme case, this behavior means that the Sm–Nd system can be used to effectively ‘see’ through many crustal processes and can provide information on the nature of the source of the rocks in question (DePaolo 1988). For voluminous crustal rocks such as granites this provides a potentially powerful proxy in constraining the nature of the various crustal blocks the granites occur within, i.e. in effect broadly mapping the crust and thus timing of crustal growth as demonstrated by Bennett and DePaolo (1987).

4 Model Ages and Residence Ages

Although initial ratios and related measures like ε_Nd are useful when comparing rocks of similar ages, the time factor implicit in Eqs. (2) and (3) makes comparison of rocks of different ages problematic. This problem can be overcome by calculating model ages, ages when the measured rock was last in isotopic equilibrium with the modelled reservoir from which it was extracted. Model ages, if calculated from the same model, are the simplest way to spatially compare isotopic data from rocks of different ages in the Sm–Nd system. Model ages determined from the Sm–Nd systems (Nd model ages) approximate the average age of continental crust in a province (McCulloch and Wasserburg 1978; DePaolo 1981, 1988; Farmer and DePaolo 1983, 1984; Liew and McCulloch 1985; Bennett and DePaolo 1987; McCulloch 1987; Fig. 2) due to the differing time-integrated behavior of Nd isotopic signatures in mantle and crustal reservoirs (Fig. 2), and ‘minimal’ changes of Sm/Nd during many crustal processes. Hence, Nd model ages estimate the time when a rock was separated from its modelled source, typically depleted mantle. This approach is most useful for magmatic rocks, but has been used for all rock types (McCulloch and Wasserburg 1978).

Although historically model ages were calculated assuming a chondritic mantle (T_CHUR), after recognition that depleted upper mantle is the dominant reservoir from which crust is extracted, most model ages now are calculated assuming depleted mantle as the source (T_DM: DePaolo 1981, 1988). Several models have been developed to model the evolution of depleted mantle. These include a model that assumes an increasingly depleted mantle (DePaolo 1981), and a linear depletion model, used herein, which assumes linear depletion from ε_Nd = 0 at ~4.56 Ga to + 10 today (Fig. 2). McCulloch (1987) also developed a linear model with assumed depletion commencing at 2.75 Ga. Recent research suggests that mantle older than ca. 3.5–3.7 Ga may have been chondritic (e.g., Hiess and Bennett 2016), and so a model similar to that of McCulloch (1987) but with depletion commencing at ca. 3.5–3.7 Ga may be more valid. This is a controversial area, however, of active research. It has been largely based on Lu–Hf studies, and it is apparent that there may have been decoupling of Sm–Nd from Lu–Hf isotopic systems (e.g., Vervoort 2011; Vervoort et al. 2019, 2020), with depletion commencing much earlier for Sm–Nd. This apparent decoupling may reflect disturbance of the Sm–Nd system (e.g., Hammerli and Kemp 2021). Until this matter is resolved, we still use the linear depletion model back to 4.5 Ga as our preferred model for the depleted mantle growth curve and for calculating Nd model ages. The linear depletion model results in older calculated model ages. Given that in isotopic mapping it is the relative spatial changes in model ages that are of more concern than absolute ages, the choice of model is not expected to have a significant effect on results and conclusions derived from isotopic mapping.

Model ages can also be determined using single stage or multi-stage (mostly two-stage) models. Single stage models (Fig. 2) infer that the main change in Sm/Nd occurred during mantle extraction and that crustal processes (e.g., fractional crystallization, contamination, alteration) have not significantly modified this ratio. Although geochemically unrealistic (see variations in Sm/Nd of crustal reservoirs in Table 1), single stage models were a feature of many early studies and did provide useful results (e.g., Bennett and DePaolo 1987). As single stage model ages (T_DM) become increasingly prone to error with increasing Sm/Nd when sample evolution curves become sub-parallel with mantle evolution curves, model ages should not be calculated for ¹⁴⁷Sm/¹⁴⁴Nd ratios over 1.4–1.5. This is generally not a problem as most felsic igneous rocks have ¹⁴⁷Sm/¹⁴⁴Nd ratios between 0.09 and 0.12 (e.g., Sun et al. 1995).

Two-stage model ages (denoted as T_2DM) are commonly used for felsic igneous rocks (Liew and McCulloch 1985) to correct for changes in Sm/Nd ratios due to partial melting, fractional crystallization, magma mixing, contamination or hydrothermal alteration. Two-stage models use the measured ¹⁴⁷Sm/¹⁴⁴Nd ratio to model evolution back in time to the independently-determined magmatic age of the rock. From the magmatic age to the modelled extraction age, an assumed ¹⁴⁷Sm/¹⁴⁴Nd ratio is used to calculate the sample evolution (Fig. 2). In this review, a ¹⁴⁷Sm/¹⁴⁴Nd ratio of 0.11, which is based on the average upper continental Sm and Nd concentrations reported by Rudnick and Gao (2003), was used. For a given sample or study, T_2DM may be younger or older than T_DM, depending on the assumed age and the measured and assumed ¹⁴⁷Sm/¹⁴⁴Nd ratios used. Two-stage model ages provide more consistent ages on a regional basis (Champion and Cassidy 2008), and allow ages to be determined for samples with high measured ¹⁴⁷Sm/¹⁴⁴Nd.

Residence ages (T_Res), which model age differences between the modelled crustal extraction from the mantle and melting and crystallization of the sample, can also be used as a variable in isotopic maps, and are calculated using the equation:

$${\text{T}}_{{\text{Res}}} = {\text{T}}_{2{\text{DM}}} -{\text{T}}_{{\text{magmatic}}}$$

(4)

Residence ages provide an indication of the time between when the granite source was extracted and when the source was reworked (melted) to produce the granite. Granite formed by the reworking of largely juvenile (i.e., young) crust will have younger residence ages than granites formed from older crust. Combining T_2DM and T_Res with geochronological data such as magmatic and inherited zircon ages provides constraints on the age of crust in a region, the age of crustal reworking and melting, and on the protolith of individual granite units.

It must be stressed that model and residence ages are just expression of a model and that absolute ages differ between models. As discussed by many workers (DePaolo 1988; Kemp et al. 2009; Vervoort and Kemp 2016) processes that produce the isotopic signature of a rock are complex, and Sm–Nd model age calculations contain a number of assumptions, as discussed above and outlined in detail by Champion and Huston (2016) and Vervoort and Kemp (2016). As an example, models ages for felsic igneous rocks, if taken at face value, imply that the source of the rock in question was homogeneous and formed in a single event, assumptions that are rarely true. Moreover, even when these criteria are met, a model age is still only an approximation with many uncertainties. There is abundant evidence of complex sources as the mixed role of crustal, juvenile and assimilated components has been demonstrated by many studies (Kemp et al. 2007, 2009; Fisher et al. 2017). In such complex open systems the calculated Nd model ages are best interpreted as average ages of all preceding crustal growth events. Importantly, as discussed by Champion and Huston (2016) it is the regional variations in model ages that are most applicable and not the absolute ages. These relative variations highlight regional changes in mantle flux, crustal growth and geodynamic environments. Moreover, use of isotopic data in a regional manner, especially in regions with a significant range of ages, tends to smooth out and downplay these ‘secondary’ effects. As shown by Champion and Huston (2016) this is evident when the map scale is taken into account. Small-scale images, covering a large region (1000+ km), exhibit significantly less scatter of data (less noise) than larger scale images (100’s of km). This needs to be considered when using isotopic maps, particularly when using large scale maps. For metallogenic analysis it is the continent to regional scale isotopic maps that are of most use.

5 Mineral Systems and Spatial and Temporal Variations in Sm–Nd Isotopic Signatures

The advantage of a mineral systems approach is that increased knowledge of the 4D (space and time) evolution of a geological terrane (i.e. regions with mutual tectonic histories), effectively allows for a better understanding of the metallogeny of that terrane, and by extension, more effective mineral exploration (McCuaig et al. 2010; Huston et al. 2016). This includes gaining a greater knowledge of the lithosphere of a terrane, its architecture, and its evolution in space and time (Fig. 1b). Radiogenic isotopes are very applicable for this purpose for geochronology and, as tracers, are able to constraint the nature, composition and evolution of the lithosphere. Examples of this approach, using Sm–Nd isotopic data coupled with other geochemical and geochronological data sets from felsic igneous rocks to help constrain crustal evolution, include the western United States (e.g., Zartman 1974; Farmer and DePaolo 1983, 1984; Bennett and DePaolo 1987; Wooden et al. 1998), and more recently Australia (Champion and Huston 2016). As granites are mainly derived from the lower and middle continental crust and may have volumes of X0 to X00 km³, these rocks can constrain the evolution of crustal growth. Studies of granites and related rocks provide indirect constraints on the origin of the crustal domains that host these rocks (Farmer and DePaolo 1983, 1984).

As indicated above, most information from isotopic data can be obtained by assessing spatial and secular changes of isotopic signatures (Fig. 3). Temporal variation can be included by time slices (e.g., Mole et al. 2014; Champion and Huston 2016; Osei et al. 2021), or by using time-independent isotopic variables such as model ages. Documenting geographic changes in isotopic parameters relies on data classification. Rapid and objective visualization of regional isotopic data, such as Nd model ages, is best produced using computer-assisted imaging, including interpolated images, where data are grouped and displayed as classes (Fig. 4). Figure 4 shows regional changes in Nd model ages that correspond well to patterns deduced from the isotope data alone (Fig. 3) and constrains crustal growth in the Pilbara Craton. The results of interpolation will depend to some extent on the interpolation process used and the technique used to bin interpolated results (i.e., identify intervals using equal intervals, percentiles, natural breaks etc.), both of which will affect the produced image. After considering interpolation techniques (Slocum et al. 2009), Champion and Huston (2016) concluded that natural neighbor classification using natural breaks in data values worked best for Sm–Nd model age images, and produced results akin to those produced by manual contouring.

Fig. 3

Two-stage depleted mantle model ages (a) and ε_Nd (b) versus magmatic ages for the Pilbara Craton, Western Australia (modified from Champion and Huston 2016). Isotopic data points are coloured by geological province: East Pilbara Terrane versus the Karratha, Regal and Sholl terranes of the West Pilbara Superterrane (see Fig. 4). The East Pilbara Terrane shows decreasing ε_Nd and approximately constant T_2DM with decreasing magmatic age (Type 1 behaviour in Table 2), consistent with reworking of old crust with minimal juvenile involvement. The exceptions to this are 2.95–2.85 Ga magmatism in the western part of the East Pilbara Terrane, which include a more significant juvenile component (younger T_2DM, more positive ε_Nd), probably related to crustal growth of the West Pilbara Superterrane (Champion and Smithies 2000; Smithies et al. 2005). Data shown spatially in Fig. 4. Data sources as reported in Champion (2013). Modified after Champion and Huston (2016)

Fig. 4

Gridded Nd two-stage depleted mantle model age (T_2DM) map for the Pilbara Craton, Western Australia. Isotopic data used to create the grid are also shown (data and data sources as reported in Champion 2013). Also shown are: the boundary between the Karratha & Regal Terranes and the Sholl Terrane (all three comprise the West Pilbara Superterrane); the boundary between the West Pilbara Superterrane and East Pilbara Terrane and the boundary between the East Pilbara Terrane and the Kurrana Terrane (nomenclature follows van Kranendonk et al. 2007). Grid colours in areas with no samples are based on interpolation and may have no relationship with underlying deeper crust. Modified after Champion and Huston (2016)

Absolute values of isotopic parameters such as model ages are less useful than geographical and/or secular variations. Champion and Huston (2016) illustrated interpretation of such maps, as well as their advantages and disadvantages. In addition to problems such as data smoothing/averaging, and areas of no data and artefacts of interpolation, a consistent difficulty concerns the ambiguity of the isotopic data itself with a variety of possible interpretations. These images must be used with the original data (ε_Nd-time and T_2DM-time plots), to fully understand the isotopic signatures and complexities within each block, and with other geological and geophysical data, including data and images from other isotopic systems (U–Th–Pb: Huston and Champion 2023; Lu–Hf: Osei et al. 2021).

6 Using Radiogenic Isotope Maps and Mineral Systems: Specific Examples

Although radiogenic isotopes have been used in many ways during metallogenic studies, it is the continent to regional scale isotopic maps that are of interest here. Maps at these scales can be used to identify large-scale components and processes from a variety of mineral systems (Champion and Huston, 2016):

• Determining relationships between mineral systems and isotopic domains. Examples include: orogenic gold, volcanic-hosted massive sulfide (VHMS) base-metal and komatiite-associated nickel deposits (KANS) in the Yilgarn Craton, Western Australia (Cassidy and Champion 2004; Cassidy et al. 2005; Huston et al. 2005; Mole et al. 2013, 2014; Champion and Huston 2016; Osei et al. 2021); porphyry Cu and Mo deposits in the western United States (Zartman 1974) and in the Central Asian Orogenic Belt (Wu et al. 2021); and Proterozoic IOCG belts in Australia (Skirrow 2013). These empirical correlations can be extracted, tested and applied as predictive tools. Osei et al. (2021) provide a good example of extracting and testing correlations between isotopic signatures and mineral deposits for the Yilgarn Craton. Similarly, Huston et al. (2014) show how correlations observed initially in the Yilgarn Craton were applicable for Archean and Proterozoic VHMS deposits, in general.

• Identifying cratonic blocks that may have greater metal endowment and have other favorable metallogenic characteristics, (for example, regions of thicker lithosphere that focus mantle melts around their margins (Begg et al. 2010; Mole et al. 2014). Recently, Hoggard et al. (2020) demonstrated that margins of thicker lithosphere coincide with the location of the world’s sediment-hosted metal deposits, and suggested a role for long-lived lithospheric edge stability. Such old stable margins should be evident in isotopic maps and Huston et al. (2020), using Pb isotopes, show that such a relationship does exist. Old crustal blocks and their underlying mantle lithosphere may also have been the foci of (repeated) mantle melts that provided metal sources for later reworking (e.g., Groves et al. 2010). Similarly, Wu et al. (2021) showed a relationship between crustal thickness and isotopic signature in the Central Asian Orogenic Belt and that the relative distribution of porphyry Cu–Au and porphyry Cu–Mo deposits was related to both isotopic signature and crustal thickness, with porphyry Cu–Au deposits in the more isotopically juvenile thin crust zones.

• Establishing old continental margins, particularly accretionary margins that are favorable sites for porphyry copper and related deposits, especially where accompanied by juvenile isotopic signatures (Champion 2013; Wu et al. 2021). This also includes the ability to identify and map continental fragments within such accretionary orogens. For example, Wu et al. (2021) were able to use Sm–Nd isotopic mapping to map the extent of cratons, and locations of microcontinents, and juvenile crustal blocks. Kemp et al. (2009, 2020) and Champion and Huston (2016) identified interpreted island arc fragments in eastern Australia based on juvenile isotopic signatures. These fragments may be important for later mineral systems as accretionary orogens are can involve lithospheric metasomatism and provide metal sources for later reworking (Groves et al. 2010).

• Identifying juvenile zones indicative of extension and/or rifting and associated deposits, such as VHMS (Huston et al. 2014) or deposits formed in primitive arc crust (e.g., porphyry Cu–Au: Champion and Huston 2016; Wu et al. 2021).

• Recognizing crustal breaks that represent major faults and sutures, or were pathways for fluids and magmas (Wooden et al. 1998). Such breaks also may delineate boundaries of metallogenic terranes (VHMS deposits, Huston et al. 2014; porphyry deposits, Zartman 1974).

• Constructing baseline maps that determine spatial or temporal zones with high magmatic, especially mantle, flux and which have been overprinted by younger crustal growth.

Many of the above act in combination and repeatedly and are discussed below to highlight steps involved proceeding from conceptual mineral systems to exploration targeting (McCuaig et al. 2010).

6.1 Regional Samarium–Neodymium Isotopic Signatures in the Yilgarn Craton and Their Relationship to Nickel, Copper–Zinc–Lead and Gold Deposits

In the Yilgarn Craton, features of the regional Sm–Nd signature of granites of the Yilgarn Craton correspond with a range of mineralization styles, including KANS, VHMS and orogenic gold deposits (Fig. 5; Cassidy et al. 2002, 2005; Huston et al. 2005, 2014; Champion and Cassidy 2008). Osei et al. (2021), using both Sm–Nd and Lu–Hf isotopes with a more rigorous statistical approach, showed that KANS deposits are spatially associated with terranes with pre-existing evolved crust, as identified by Nd T_2DM model ages (Cassidy et al. 2005; Fig. 5). Osei et al. (2021) refined this relationship and showed that these deposits were located along the edge of, and not within, older cratonic blocks. This relationship holds despite variations in komatiite type and age. Cassidy et al. (2005) inferred a relationship between orogenic gold deposits and older crustal terranes, again updated by Osei et al. (2021) to show a more complex relationship although still located along edges of older crust. There is also a global antithetic relationship between KANS and VHMS deposits (Groves and Batt 1984; Cassidy et al. 2005). Huston et al. (2005, 2014) and Osei et al. (2021) showed that VHMS deposits in the Yilgarn Craton are associated with regions with isotopically primitive crust (Nd model ages close to magmatic ages, i.e. with young residence (T_Res) ages). The latter are age-independent, and readily identify young (juvenile) versus older crustal domains. In regions where felsic magmatic ages have small ranges, T_Res maps are very similar to T_2DM maps. Either works well and T_Res maps of the Yilgarn Craton also identify juvenile zones and their spatial association with VHMS deposits.

Fig. 5

a Location of gold, and b nickel (nickel sulphide) and volcanic hosted massive sulphide (Pb–Zn) deposits, by size, in the Yilgarn Craton, superimposed over gridded two-stage depleted mantle model age (T_2DM) map. Image constructed from 305 data points. Data and data sources as reported in Champion (2013), from which the figure was modified. Grid colours in areas with no samples are purely based on interpolation and may have no relationship to the underlying crust. Mineral deposit locations are from Geoscience Australia (2022). Yilgarn Craton terrane boundaries from Cassidy et al. (2006). WAE = Western Australian Element, CAE = Central Australian Element. Element boundaries as in Huston et al. (2012). Refer to Osei et al. (2021) for more updated images of the same region. Modified after Champion and Huston (2016)

Reasons for the Yilgarn isotope-deposit correspondence are not well understood, particularly the linkage of nickel and gold deposits to isotopically more evolved crust. In part, this reflects uncertainty over the interpretation of the crustal isotopic signature, i.e., is it a direct signature of the mineral system or simply is it a proxy for some other important feature. Subsequent work (Huston et al. 2014; Osei et al. 2021) inferred both, with VHMS deposits reflecting the former and KANS deposits the latter.

Reasons for the association of gold endowment with isotopic domains of intermediate T_2DM are more enigmatic, and several mechanisms have been proposed (Cassidy et al. 2005; Osei et al. 2021). Blewett et al. (2010), Czarnota et al. (2010) and McCuaig et al. (2010) argued that the Nd model age map identified lithospheric-scale architecture adjacent to paleocraton margins. This is not dissimilar to models, such as those of Goldfarb and Santosh (2014), which suggested deep (possibly slab-related) ore-forming fluids in the Phanerozoic gold deposits of Jiaodong Province, China, were channeled into the upper crust along continental-scale fault systems. Such faults occur proximal to old craton margins, readily identifiable by isotope systems such as Sm–Nd. As outlined by Osei et al. (2021) these lithospheric structures may also be controlling location of magmas, such as sanukitoids and lamprophyres, providing a possible link between location of gold deposits and the magmas and magma pathways that may have carried the gold into the crust (e.g., Beakhouse 2007). A recent example of this possible sanukitoid association are the recently discovered Hemi gold deposit in the Pilbara Craton, Australia (De Grey Mining Ltd 2022). This, and associated deposits, occur closely associated with a belt of ca. 2.945 Ga sanukitoid intrusions in the Mallina Basin, southwest of Port Hedland, in an isotopically more juvenile zone, that approximates the Sholl Terrane but also extends to within the westernmost part of the neighboring older East Pilbara Terrane (Smithies and Champion 2000; Figs. 3, 4).

6.1.1 Isotopic Domains and Komatiite-Associated Nickel Deposits: Control By Lithospheric Architecture?

The distribution of KANS deposits in Yilgarn crustal domains has been assessed by many authors. Barnes and Fiorentini (2010a, b, 2012) found that in the Yilgarn, nickel endowment in the Yilgarn was concentrated in the Kalgoorlie Terrane, suggesting that this concentration resulted from factors that enabled high and prolonged fluxes of komatiitic magmas into the Kalgoorlie Terrane, following the craton-margin model of Begg et al. (2010; Fig. 6). This model suggests that komatiitic melts from upwelling mantle plumes were channeled away from regions of thicker lithosphere to areas of thinner lithosphere with accompanying decompression melting. The Yilgarn Nd isotopic map (Champion and Cassidy 2007, 2008; Mole et al. 2013; Osei et al. 2021) is consistent with this model with as it clearly identifies a break between two lithospheric blocks that represents an old continental margin (Begg et al. 2010) and/or marginal basin (Krapez et al. 2000). Mole et al. (2014) used Lu–Hf data from inherited zircons, along with their U–Pb ages, to show that similar architecture may have also controlled older komatiitic magmatic events and associated KANS deposits.

Fig. 6

The generalised model of Begg et al. (2010) illustrates preferential flow of impinging mantle plume towards, and subsequent localisation of nickel sulfide deposits within, thinner lithosphere along the margins of thick (> 150 km) lithospheric blocks, e.g. with intervening marginal basin (a) or along an old paleo-margin (b). Either model may be applicable for the Yilgarn Craton (Fig. 5). Modified after Begg et al. (2010)

Crustal isotopic maps define lithospheric architecture. Begg et al. (2010) summarized several mechanisms whereby lithospheric architecture, particularly thinner lithosphere along craton margins, would be more favorable for KANS and related deposits. These mechanisms include the preferred flow of mantle plumes and accompanying partial melts into such regions, and passage of such melts into the crust along large-scale fault systems. Isotopic systems identify favorable regions at a craton- to district-scale, whereas other factors, such as the availability of sulfur, control mineralization at the district-scale (Fiorentini et al. 2012).

Samarium–Nd isotopic maps (e.g., Fig. 5) clearly identify old continental margins, and can be used in exploration targeting as proxies to recognize such architecture. For example, the Nd isotopic map of the Pilbara (Fig. 4) identifies old cratonic margins, and identification of old margins is important for a number of other mineral systems.

There are other possible explanations for the correlation between the Nd isotopic data and nickel mineralization. McCuaig et al. (2010) and Fiorentini et al. (2012) highlighted other proxies, such as the location of felsic volcanic rocks and VHMS deposits, to identify extensional zones, which, like old continental margins, are another active pathway for komatiite lava. This is consistent with Osei et al. (2021) who demonstrate that Ni–Cu–PGE mineralization is located along the margins of older crust. Juvenile isotopic zones also allow recognition of major extensional zones that are evident within the Yilgarn Craton, especially in the Eastern Goldfields Superterrane. These juvenile isotopic zones are bordered by more isotopically evolved zones, and it could be inferred that KANS mineral systems are closely spatially associated with second order extensional zones. It is probable that both mechanisms were operative in the Yilgarn Craton, albeit at different scales.

6.1.2 Isotopic Domains and Volcanic-Hosted Massive Sulfide Deposits

Huston et al. (2005, 2014), and subsequently Osei et al. (2021), showed an empirical relationship between the distribution of VHMS deposits in the Yilgarn Craton and more juvenile crust, as identified by younger T_2DM (Fig. 7). Huston et al. (2014) also showed a similar relationship with Pb isotopes in galena. They used the close correspondence between the Nd isotopes in the granites and Pb isotope signatures in ore minerals to suggest that the juvenile Nd signature reflected a critical aspect of the VHMS mineral system. More, importantly, they suggested this signature could be used predictively for VHMS deposits in exploration targeting. Huston et al. (2014) investigated this further and demonstrated a strong relationship between endowment and isotopic signature, at least for Archean (and, possibly, Proterozoic) VHMS provinces around the world. The association of VHMS deposits with juvenile isotopic zones strongly suggests a link between juvenile crustal growth and VHMS mineralization. Neodymium and Pb data in highly endowed domains indicate limited interaction with pre-existing crust. Huston et al. (2014) argued that isotope data indicate a favorable setting for VHMS deposits in extensional zones that are characterized by high-temperature juvenile magmas and extensive structuring.

Fig. 7

Plot of Cu, Zn, Pb and combined Cu–Pb–Zn metal endowment (in tonnes per km²) for volcanic-hosted massive sulfide deposits in Archean cratonic blocks in Canada and Australia, highlighting the much greater endowment in blocks with a juvenile isotopic signature (Whundo, Teutonic, Cue and Abitib-Wawa). Endowment figures from Table 3 of Huston et al. (2014), which were updated from Franklin et al. (2005). Modified after Champion and Huston (2016)

6.2 Iron Oxide–Copper–Gold Deposits and Isotopic Gradients: Mapping Old Continental Margins?

Skirrow (2013) showed that many iron-oxide copper–gold (IOCG) provinces are associated with isotopic gradients (Fig. 8), a relationship observed at continental and regional scales. These gradients may record an earlier accretionary continental margin. The isotopic characteristics become increasingly juvenile toward the continental margin, as seen in continental margins such as in the southwestern USA (Bennett and DePaolo 1987) and the Tasman Orogen of eastern Australia (Kemp et al. 2007; Champion et al. 2010). An accretionary margin has been inferred for the IOCG provinces (Skirrow 2013). These include at 1.85 Ga and possibly ca 1.95 Ga for the Mount Isa region (Korsch et al. 2011; McDonald et al. 1997; Gregory et al. 2008), ages of ca. 1880–1650 Ma for the southern Arunta region (Zhao and McCulloch 1995; Scrimgeour et al. 2005), and 1.85 Ga and possibly 2.5 Ga for the Gawler Craton at (Swain et al. 2005; Payne et al. 2009; Korsch et al. 2011;). As noted by Champion and Huston (2016), the isotopic data for the southern part of the Northern Territory becomes increasingly primitive southward, consistent with a relatively long-lived or episodic accretionary margin (Fig. 9).

Fig. 8

Plot of iron oxide copper gold (IOCG) provinces overlain over gridded two-stage depleted mantle model age (T_2DM) map of Australia (Champion 2013). Updated from Skirrow (2013). As noted by Skirrow (2013), there is a good correlation between these mineral provinces and apparent gradients (breaks) in the regional T_2DM map. Refer to Champion (2013) for Sm–Nd data sources. Element nomenclature follows Huston et al. (2012)

Fig. 9

Two-stage depleted mantle model (T_2DM) ages versus latitude for felsic igneous rocks of the Northern Territory (modified after Champion and Huston 2016). Note the general decrease in minimum and max T_2DM with decreasing latitude, highlighted by gray area and black arrow. This is particularly evident south of latitude 20 degrees. Figure modified from Champion (2013); refer to that reference for Sm–Nd data and data sources

The association with old margins has been inferred by others (e.g., Skirrow 2010), based on seismic reflection data for the Olympic Dam IOCG deposit. Groves et al. (2010) interpreted that large IOCG deposits were located in intracratonic settings but close to lithospheric boundaries indicated by regional isotopic patterns. Groves et al. (2010) also suggested that larger IOCGs formed shortly after (within 100–200 Myr) supercontinent formation, indicating that age data can assist in identifying potential IOCG provinces. This relationship, however, in part reflects classification of IOCG deposits. Inferred IOCG deposits at Tennant Creek, Northern Territory, for example, which Groves et al. (2010) interpreted as high grade Au (+Cu) deposits, not true IOCGs, were generated during formation of the Nuna supercontinent (Huston et al. 2012).

Given the association of IOCG deposits with craton margins/sutures, Groves et al. (2010) postulated that subduction-related metasomatism of lithospheric mantle may have been an important element in the IOCG mineral system. Partial melts from the mantle lithosphere transported copper, gold and volatiles into the crust, although local country rocks may have supplied uranium (Skirrow et al. 2007). These lithospheric melts would have been oxidized (e.g., Rowe et al. 2009) with enhanced capacity to carry copper and gold (e.g., Jégo et al. 2010; Loucks 2014). The presence of igneous rocks derived from metasomatized lithosphere combined with regions of pronounced gradients in isotopic maps may highlight zones with IOCG mineral potential. Melts from metasomatized lithosphere would not necessarily have juvenile isotopic signatures as the timing of metasomatism and the presence of sediments in the subduction component responsible for metasomatism may produce evolved isotopic signatures.

For example, Tasmanian Jurassic dolerites in Tasmania have evolved signatures (ε_Nd ~ −6) even though these rocks were interpreted as being derived from subduction-related metasomatized mantle lithosphere (Hergt et al. 1989). More recently, Lu et al. (2013) have highlighted isotopically evolved mafic magmas associated with porphyry Cu mineralization in the Western Yangtze Craton, China, which they suggested were derived from ancient metasomatized lithospheric mantle.

Johnson and McCulloch (1995) and Skirrow et al. (2007) suggested, based on correlations between Sm–Nd isotope signatures and Cu contents, that juvenile mantle melts sourced copper at the giant Olympic Dam deposit, although no compelling evidence exists that these are lithosphere melts. Arndt (2013) points out that large volumes of melt are unlikely to be generated from such sources. This does not, however, change the apparent relationship between Nd (and Pb) isotopic signatures and IOCG belts as pointed out by Skirrow (2013; Fig. 8).

Isotopic data can also be visualized with maps of crustal residence (T_Res). This approach can be applied to the Gawler Craton where widespread granites have similar or slightly older ages (ca. 1860–1570 Ma) than the accepted age of IOCG mineralization (ca. 1600–1570 Ma; Skirrow et al. 2007). Hence, the crustal residence map (Fig. 10) provides an image of the characteristics of Gawler crust during IOCG mineralization. Figure 10 shows that isotopic data support the presence of continental margins on both sides of the Gawler Craton. This relationship is consistent with geological relationships indicative of a convergent margin setting on the southwestern part of the Gawler Craton at ca. 1630–1610 Ma and ca. 1700 Ma (Swain et al. 2008; Ferris and Schwartz 2004; Payne et al. 2010). This raises the question of why IOCG mineralization is localized on the eastern and northeastern margins of the Gawler Craton and not elsewhere in the craton, and suggests additional controls on the location of mineralization. For example, mineralization may in part reflect focusing of mantle-derived melts into zones with thinner lithosphere. Skirrow (2010) suggested a model for the Olympic Dam Province in which delamination resulted in mafic and ultramafic magmatism that was focused into, the delaminated region, which was ultimately responsible for mineralization. There is perhaps evidence for such a mantle-derived magmatic flux in the Sm–Nd isotopic data, best seen through specific time slice isotopic plots.

Fig. 10

Crustal residence age map for the Gawler Craton, based on T_Res of ca. 2000–1550 Ma granites from that craton. Outlines of Gawler and Curnamona Cratons and the geological domains of the Gawler Craton are from Ferris et al. (2002). Refer to Champion (2013) for relevant data and data sources. Modified after Champion and Huston (2016)

6.2.1 Refining Iron Oxide–Copper-Gold Search Space—Isotopic Time Slice Diagrams

Time slice Nd model age maps effectively provide an isotopic snap shot of the crust at the time of any specific geological event, such as IOCG mineralization. Such plots can be simply constructed. For example, for 1800 Ma the model age is simply calculated as follows:

$${\text{T}}_{2{\text{DM}}(1800\;{\text{Ma}})} = {\text{T}}_{2{\text{DM}}} -1800;\;{\text{for}}\;{\text{all}}\;{\text{T}}_{2{\text{DM}}} > 1800\;{\text{Ma}}$$

(5)

In provinces with a limited age range for magmatism, this will produce similar images to T_Res maps. In provinces with a large range of magmatic ages, maps for pseudo-time slices based on Nd model ages can be produced (Fig. 11).

Fig. 11

Model age time slice figures (A: at 2500 Ma; B: at 1800 Ma; C: at 1550 Ma) for the Gawler and Curnamona Cratons of the South Australian Element, Australia. The Olympic IOCG field (named after the Olympic Dam IOCG deposit; from Skirrow 2013) is denoted by orange dashed line in A and C. Note the persistent more juvenile isotopic embayment located around the Olympic Dam IOCG deposit (near the word ‘Olympic’ on a, b and c). Refer to Champion (2013) for Sm–Nd data and data sources. Modified after Champion and Huston (2016)

The Gawler Craton is particularly amenable to this approach as felsic magmatism spanned over 1500 Myr (Reid and Hand 2012). Figure 11 illustrates 2500, 1800 and 1550 Ma time slices, corresponding to the Archean, intermediate (pre-mineralization) and immediate post-mineralization craton structure. The Olympic Dam IOCG province straddles the margin of the 2500 Ma and older crust, and in the younger time slices, the isotopic data are zoned along this margin, with an old crustal block in the south and the juvenile embayment in the north. This embayment coincides with the Olympic Dam deposit (Fig. 11c). Although this zone could represent a long-lived juvenile feature that dates back to the Archean, it is more likely that it reflects juvenile post-Archean crustal growth close to the time of mineralization (Skirrow et al. 2007), consistent with a juvenile source of copper (Johnson and McCulloch 1995).

Further evidence for contemporaneous mantle input lies in the felsic rocks of the Hiltaba Suite magmatism (ca. 1595–1570 Ma) which become increasingly more A-type toward the deposit (Budd 2006), consistent with a greater mantle input, towards the deposit (Budd 2006). Metal could be derived from contemporaneous mafic/ultramafic magmatism through magmatic-hydrothermal processes or leached from such pre-existing rocks (Campbell et al. 1998). Both processes are consistent with the model of Groves et al. (2010) if the mafic/ultramafic magmatism was derived by partial melting of the lithosphere and at least partly consistent with the Begg et al. (2010) model, in which mafic–ultramafic magmatism was focused into the region around Olympic Dam along a crustal break. The 1550 Ma time slice (Fig. 11c) approximates crustal features during Hiltaba magmatism, and IOCG formation. More importantly, the location of a more juvenile isotopic zone along the isotopically defined paleo-continental suture may be highlighting a more favorable exploration target.

6.3 Predictive Analysis Based on Radiogenic Signatures: Granite-Related Mineralization

Ishihara (1977, 1981) first recognized that the redox state of granite magmas appeared to have some control over the associated metallogenesis. These studies were expanded to show that the degree of chemical evolution of the granite magma also controls metallogenesis (Blevin and Chappell 1992; Blevin et al. 1996; Thompson et al. 1999; Lang et al. 2000; Blevin 2004; Fig. 12). Thompson et al. (1999) showed that rare metal (Sn, Mo and W) deposits are associated with continental material in backarc or post/non-arc extensional environments, whereas copper ± gold deposits are related to more primitive crust arcs, consistent with the oxidized nature of arc rocks (Parkinson and Arculus 1999). More recently, Richards (2011) and Loucks (2014) have suggested the important additional role of water in arc-related magmas for copper mineralization.

Fig. 12

a Rb/Sr ratio versus Fe₂O₃/FeO ratio plot modified after Blevin et al. (1996). The plot illustrates the relationship between the degree of oxidation and compositional evolution of the magma (based on whole-rock compositions) and the dominant commodities in related mineralisation. Intrusion related gold deposit (IRGD) field from Blevin (2004). b Plot of oxygen fugacity versus total Fe concentration in the magma, also showing tectonic setting. Plot modified after Thompson et al. (1999) and Lang et al. (2000)

These relationships suggest that Sm–Nd (and other) isotopic data can delineate prospective zones for porphyry copper and copper–gold deposits, by identifying regions with more juvenile isotopic signatures. In the Macquarie Arc in New South Wales, Australia, latest Ordovician–earliest Silurian porphyry copper–gold deposits are associated with isotopically juvenile magmatism (Cooke et al. 2007; Fig. 13). This domain is well-mapped by regional- and national-scale T_2DM and T_Res images (Fig. 13), despite the limited number of data points in the data set. The T_Res map also highlight isotopically juvenile areas in the northern New England Orogen, which hosts known copper–gold deposits such as the historically important Mount Morgan deposit in central Queensland. Mount Morgan is associated with rocks interpreted to be related to either a primitive continental arc (Morand 1993) or island arc (Murray and Blake 2005). Although magmatic rocks associated with these copper–gold deposits are isotopically primitive, they are much better discriminated using T_Res than T_2DM images. The former image better identifies potentially mineralized units, as they are focusing on zones with magmatic ages very close to model ages. Regardless, it is evident that both isotopic maps (T_2DM and T_Res) are indicating juvenile zones amenable to porphyry and related deposits, even where the numbers of analyzed samples is low. These regions are also identifiable in other isotopic figures, such as ε_Nd versus time (see Champion 2013).

Fig. 13

Location of Cu–Au deposits superimposed on the gridded Nd residence age map for the Tasman Element and surrounding regions (South (SAE) and North (NAE) Australian elements). Copper–gold deposits (from Geoscience Australia’s Australian Mines Atlas, downloadable via their portal https://portal.ga.gov.au/persona/minesatlas) are shown as blue circles. Also shown are the Delamerian, Lachlan, Thomson and Mossman orogens of the Tasman Element. There is a good correlation between many copper–gold deposits and zones with young residence ages, notably in the central Lachlan Orogen and the central New England Orogen. Copper–gold deposits outside of these two zones are almost exclusively not magmatic-related. Location of Nd samples used to create the grid are shown as black circles. Data and data sources are given in Champion (2013), after which the figure is modified. Boundaries of Delamerian, Lachlan, Thomson, Mossman and New England (NEO) orogens modified from Stewart et al. (2013)

Recently, Wu et al. (2021) demonstrated similar relationships between Nd isotopic signatures and maps and the many porphyry Cu–Au and Cu–Mo deposits in the Central Asian Orogenic Belt (CAOB). Using an extensive Sm–Nd isotopic dataset, they were able to map locations and extents of cratons, as well as microcontinents and juvenile crustal blocks. Wu et al. (2021) were able to show that the porphyry deposits were largely spatially located within isotopically juvenile crust, marginal to cratons or microcontinents, similar to that documented for eastern Australia. Further, consistent with Thompson et al. (1999; Fig. 12), Wu et al. (2021) showed that porphyry Cu–Mo deposits were generally found within isotopically less juvenile crust than the porphyry Cu–Au deposits. Wu et al. (2021) took their analysis further, and, after estimating crustal thickness for each deposit (based on Sr/Y and La/Yb ratios of the host granites), were able to define a negative linear correlation between εNd and crustal thickness for the porphyry deposits. In this scheme, porphyry Cu–Au deposits were associated with thinner, more isotopically juvenile crust, in contrast to porphyry Cu–Mo deposits that were associated with thicker more isotopically evolved crust. Loucks (2014), in his review of Phanerozoic porphyry mineralization, demonstrated similar results, showing that Cu–Au porphyries were characterized by generally lower Sr/Y ratios than Au-poor Cu porphyries. These results are also consistent with the general observations of Thompson et al. (1999). The isotopic data for the CAOB, and eastern Australia show that exploration search space for regions with potential for porphyry Cu and Cu–Au deposits can be effectively narrowed to regions with juvenile isotopic characteristics.

The usual caveats, of course, apply. Neodymium and other isotopic data identify potentially juvenile terranes, but do not convey any information on additional important factors, such as depth of current crustal exposure, the degree of preservation of the porphyry and related deposits, the rate of uplift, the degree of metamorphism and other factors. Porphyry copper–gold deposits are also common in more mature continental arcs although the delineation of these on the basis of their isotopic signature is more problematical. Further, as shown by Lu et al. (2013), porphyry Cu-related magmas, even along suture zones, need not have juvenile isotopic signatures, especially if ancient metasomatized lithosphere is involved in the genesis of such magmas.

6.4 Isotopic Mapping and Accretionary Tectonic Settings

Isotopic maps can be integrated with other features that characterize mineral systems, including tectonic settings. An example is identification of accretionary orogens that commonly host specific mineral systems such as those that form porphyry copper and associated deposits (Hronsky et al. 2012). Accretionary margins commonly are characterized by increasingly juvenile isotopic signatures outward and with time as the margin evolves. This gradient can be useful for regional exploration targeting as seen in data for the western United States (Farmer and DePaolo 1983; Bennett and DePaolo 1987), eastern Australia (Kemp et al. 2009; Champion et al. 2010; Fig. 8) and in the southern half of Northern Territory (Fig. 9). Collins et al. (2011) suggested that different types of orogenic systems can be discriminated. For example, dominantly accretionary (or Internal: circum-Pacific) versus dominantly collisional (or External: southern Europe) can be recognized on the basis of isotopic signatures. Although we agree with this concept in general, Champion (2013) indicate that the isotopic signatures for ‘Internal’ orogens are not unique (Table 2).

Table 2 Secular variations in T_2DM and ε_Nd in felsic magmatism and possible interpretations

Isotopic maps at various scales can be used to characterize larger scale components that are indicative of metallogenic provinces, for many mineral systems. As discussed by McCuaig et al. (2010), exploration models are predicated on the use of many geological, geochemical and geophysical datasets over a range of scales. Isotopic maps, including those based on Sm–Nd data are one of many layers to be integrated with other data sets. One example of such integration is the use of isotopic data in conjunction with crustal boundaries interpreted from geophysical data (e.g., Korsch and Doublier 2016).

7 Future Developments and In-Situ Analysis of Radiogenic Isotopes

The recent development of in-situ methods of analysis of radiogenic isotopes have led to a number of innovative approaches applicable to mineral system studies. These include in-situ Sm–Nd analysis of accessory minerals, such as apatite and other REE-rich accessory phases (Fisher et al. 2017, 2020; Hammerli and Kemp 2021), which enables assessment of inherited components within grains and allows characterization of temporal changes in the isotopic characteristics of these inherited components, particularly where independent age data are available from U–Pb analyses (Fisher et al. 2017). Much of this work is in its infancy and so regional data sets are not currently available. Good examples of what such data may eventually provide can be illustrated using the Lu–Hf isotopic system which behaves comparably to the Sm–Nd system (Vervoort and Patchett 1996; Chauvel et al. 2008). Examples of regional isotopic maps using Lu–Hf data for zircon investigating mineralization include Mole et al. (2014), Wang et al. (2016) and Osei et al. (2021).

Mole et al. (2014) determined Lu–Hf signatures and U–Pb ages of inherited and xenocrystic domains from zircons extracted from Yilgarn. Based on these data, they produced a series of temporal Lu–Hf snapshots (at 3050–2820 Ma, 2820–2720 Ma and 2720–2600 Ma). These snapshots indicated the Yilgarn Craton formed from several Archean micro-continents. These snapshots enabled documentation of the relationship between ca. 2.9 Ga and ca. 2.7 Ga komatiite deposits and paleo-crustal structures. The ca. 2.7 Ga craton margin of Mole et al. (2014) matched the one determined from Sm–Nd whole rock data (Champion and Cassidy 2007, 2008). Mole et al. (2014), however, identified an additional crustal block in the southwest part of the Yilgarn.

Techniques developed by Mole et al. (2014) are an advance on temporal images compiled from isotope model ages alone (Fig. 11), as the former can exclude complexities due to juvenile input related to younger crustal growth, magma mixing and other processes. The approach is not without problems though, in particular, the presence of inherited components in zircons derived from sedimentary rocks. For such zircons, the isotopic information they carry may not represent the crust in the general area of the host intrusive, but may have been brought in from other provinces. This is less problematic in Archean provinces, such as the Yilgarn and Pilbara cratons, as sedimentary rocks are not a large component of the geology (Champion 2013). This contrasts with younger terranes where sedimentary rocks are more abundant (e.g., eastern Australia: Kemp et al. 2009), particularly when transport distances of detrital zircons are considered (Sircombe 1999). The possible influence of sedimentary components, however, is also problematic for whole rock Nd maps (Champion, 2013).

The approach of Mole et al. (2014) provides a good example of what may also be achievable with in-situ microanalysis of accessory minerals for Sm–Nd. More importantly, as outlined by Fisher et al. (2017) and Hammerli and Kemp (2021), the combined use of in-situ Lu–Hf, Sm–Nd analysis and other isotopic systems offers a powerful way forward in deciphering isotopic signatures of components within granitic and other rocks, and, hence, making more robust isotopic maps. A number of studies (Mole et al. 2013, 2014; Osei et al. 2021) have already made use of in-situ Lu–Hf analysis combined with whole rock Sm–Nd data. On this note, given the comparable behavior and model age calculations for both Sm–Nd and Lu–Hf, there is no reason why hybrid maps using both systems comined cannot be produced (e.g., Fraser et al. 2020).

8 Links with Complementary Isotopic Systems and Geophysical Data Sets

Radiogenic isotopic studies have significant potential for understanding and constraining mineralization, not just as metal source tracers, but also, as demonstrated here, indicators of favorable geological setting, lithospheric architecture, and geodynamic environment. This contribution concentrated largely on isotopic maps based on Sm–Nd data from felsic magmatic rocks. Clearly, more can be done. Obvious first steps are to better link and integrate different isotopic systems. The U–Th–Pb system, using isotopes from ore minerals (Huston et al. 2020), provides one example. Champion and Huston (2016) discussed aspects of linking these two systems and showed that often, though not universally, the two systems convey similar information: regions with old or young Nd model ages correlate with regions of high or low μ (²³⁸U/²⁰⁴Pb), respectively. A good example of this Nd–Pb correlation is provided by Huston et al. (2014) for VHMS deposits in the Yilgarn Craton. The link between regional Nd isotopic data from granitic rocks and regional Pb isotopic data from mineralization is perhaps not surprising for granite-related mineralization, but is unexpected for non-magmatic mineral systems, for example shale-hosted Zn–Pb deposits. This linkage between the isotopic systems needs to be investigated in a more robust manner. In particular, there should also be investigation of regions where regional patterns from one system, e.g., U–Th–Pb in mineralized samples, differ from those from Sm–Nd (or Lu–Hf) in granites, as is evident, for example, in parts of the Pilbara Craton (see Champion and Huston 2016).

Further work can also extend the isotopic mapping approach here to other isotopic systems and sample media. As mineral systems commonly involve the entire lithosphere as well as the upper asthenosphere (Huston et al. 2016), the entire lithosphere, including the mantle component, should be mapped as shown by Begg et al. (2009) for Africa. Therefore, future isotope mapping should focus on characterizing the entire lithosphere at a range of scales, applying multiple isotopic systems to different rock types and minerals, and integrating the results with geological, geochemical and geophysical data.

Major recent advances have been made on geophysical mapping of the lithosphere, using data sets including passive/active seismic, magnetotellurics and other methods. Such geophysical data sets are available at many scales, including the global-scale for some datasets (Kennett and Salmon 2012; Kennett et al. 2013; Priestley and McKenzie 2013; Hoggard et al. 2020). In contrast to isotopic data sets, geophysical data are commonly collected in systematic grids. Significant additional interpretative power could be achieved in combining the two data types. This benefit works in both directions as geophysical data are a measure of the present-day situation and do not provide time resolution. Isotopic data, where integrated, will help to provide this temporal framework. As demonstrated by Begg et al. (2009), full use of geophysical data depends on integration with other data sets, including geochronological, geochemical and petrological data, including isotopic and data (O’Reilly and Griffin 2006; Griffin et al. 2013).

One isotopic system that has not seen systematic use in isotopic mapping is the Re–Os system. This system potentially can provide better constraints on growth and nature of the lithospheric mantle than isotopic systems such as Sm–Nd, Lu–Hf and U–Th–Pb (Shirey and Walker 1998; Carlson 2005). This potential is important as it provides a link between the lithospheric mantle and the crust which is more effectively characterized by the Sm–Nd and U–Th–Pb systems. Gregory et al. (2008), for example, linked copper mineralization at Mount Isa (Queensland, Australia) to metasomatized lithospheric mantle lithosphere possibly produced during subduction associated with a paleo-margin similar to that imaged by the Sm–Nd data. Model ages (including both mantle model ages and model depletion ages; T_RD) for mantle lithosphere extraction using the Re–Os system can be determined on lithospheric mantle samples (mantle nodules and the like: Pearson et al. 1995; O’Reilly and Griffin 2006). As the availability of these samples is limited, any future Re–Os mapping project must depend on proxies such as mantle-derived magmas (c.f., Gregory et al. 2008).

Mapping lithospheric mantle is important to mineral systems that involve, at some point in their evolution, metasomatism of the lithospheric mantle. Metasomatized lithospheric mantle may be an important source of ore metals or volatiles for many mineral styles, for example orthomagmatic nickel-copper-platinum group elements, IOCG, porphyry copper–gold, and thorium-REE deposits (Zhang et al. 2008; Groves et al. 2010; Hronsky et al. 2012; Griffin et al. 2013). Zhang et al. (2008) argued that formation of deposits related to flood basalts involved mantle plumes that interacted with metasomatized lithosphere and were emplaced within or marginal to crustal blocks of either Archean and/or Paleoproterozoic age. Griffin et al. (2013) extended this interpretation to involve reworking (melting) of (subduction-related) metasomatized mantle lithosphere for many other mineral systems. The role of metasomatized mantle in mineral systems, however, is controversial, and Arndt (2013) presented counter-arguments against this involvement.

If metasomatized lithosphere is important for mineral systems related to large igneous provinces and subduction, then recognition of zones of metasomatized mantle by isotopic and other data (e.g., magnetotellurics) is valuable for area selection (e.g., Hronsky et al. 2012). Even if mantle metasomatism is not an important process (Arndt 2013), isotopic maps still provide information on other processes. As discussed earlier, the topography of the lithosphere-asthenosphere boundary may be just as important for KANS and other orthomagmatic mineral systems (Begg et al. 2010). Hence, although the root causes of association with mineral systems with specific lithospheric architectures varies according to metallogenic model, datasets and images generated from Sm–Nd and other isotopic systems can identify this architecture and, thereby, fertile metallogenic terranes.

9 Conclusions

Although Sm–Nd isotope data have historically been used in metallogenic studies to stablish age and metal source, these data, presented as regional to continental isotope map, assist identification of metallogenic terranes favorable for many mineral systems. This approach is aided by large, spatially (and in some cases temporally) constrained data packages that are easily downloadable (Fraser et al. 2020). This in combination with available 2-D interpolation software, have allowed repeatable generation of isotopic maps, using various isotopic variables, at regional to continental scales, allowing for metallogenic interpretation over similarly large regions. Interpretations of these isotopic maps is enhanced by a mineral systems approach in which mineral deposits are recognized as products of geological processes that occur over a wide range of scales.

The use of isotopic data and maps strongly depends on the relative geochemical behavior of parent and daughter isotopes used and the rock types and minerals analyzed. For the Sm–Nd system detailed here, the largest fractionation is seen between depleted mantle and continental crust. This means that the Sm–Nd system can be used to effectively ‘see’ through many crustal processes to provide information on the source of granitic rocks. This includes model age calculations, which, with several assumptions, approximate of the average age of the crust. For large volume rocks like granites, the model age is an excellent proxy that constrains the character of crustal blocks that host the granite. This capability is enhanced by regional to continental scale maps constructed with Sm–Nd data for these granitic rocks. These maps can be used to identify larger-scale portions of mineral systems indicative of metallogenic tracts, for many mineral systems. Examples of uses include:

• demonstration of associations if specific mineral systems and characteristic isotopic domains;

• recognition of old and/or thick cratonic blocks;

• identification of tectonic regimes favorable for mineralization;

• mapping of isotopically juvenile zones indicative of zones of rifts or primitive arcs;

• identification of crustal breaks and possible sutures that localize mineral systems for various reasons; and

• producing baseline maps that assist in identifying regions/periods characterized by magmatic, especially mantle, input.

As exploration models currently are built with many different types of geological, geochemical and geophysical data over many scales, isotopic maps are just another layer for integration. Future work should focus on the spatial and temporal evolution of the entire lithosphere, including the lithospheric mantle, to assist in more effective mineral exploration. This must involve integration of radiogenic isotope data with other data, particularly geophysical data, which has the advantage of being able to directly image the lithosphere and being of a more continuous nature that often decidedly lumpy isotopic data sets.